Learning with Calibration: Exploring Test-Time Computing of Spatio-Temporal Forecasting

TL;DR

This paper introduces a novel test-time calibration method for spatio-temporal forecasting that corrects biases in real-time, improving accuracy without complex retraining, suitable for large-scale applications.

Contribution

The paper proposes ST-TTC, a test-time calibration framework that captures and corrects non-stationary biases efficiently during inference, bypassing intensive training procedures.

Findings

Effective bias correction in real-world datasets

Improves forecasting accuracy with low computational overhead

Demonstrates universality and flexibility across domains

Abstract

Spatio-temporal forecasting is crucial in many domains, such as transportation, meteorology, and energy. However, real-world scenarios frequently present challenges such as signal anomalies, noise, and distributional shifts. Existing solutions primarily enhance robustness by modifying network architectures or training procedures. Nevertheless, these approaches are computationally intensive and resource-demanding, especially for large-scale applications. In this paper, we explore a novel test-time computing paradigm, namely learning with calibration, ST-TTC, for spatio-temporal forecasting. Through learning with calibration, we aim to capture periodic structural biases arising from non-stationarity during the testing phase and perform real-time bias correction on predictions to improve accuracy. Specifically, we first introduce a spectral-domain calibrator with phase-amplitude modulation to mitigate periodic shift and then propose a flash updating mechanism with a streaming memory queue for efficient test-time computation. ST-TTC effectively bypasses complex training-stage techniques, offering an efficient and generalizable paradigm. Extensive experiments on real-world datasets demonstrate the effectiveness, universality, flexibility and efficiency of our proposed method.